Owens E, Blazek B
J Speech Hear Res. 1985 Sep;28(3):381-93. doi: 10.1044/jshr.2803.381.
A series of VCV nonsense syllables formed with 23 consonants and the vowels /a/, /i/, /u/, and / / was presented on videotape without sound to 5 hearing-impaired adults and 5 adults with normal hearing. The two-fold purpose was (a) to determine whether the two groups would perform the same in their identification of visemes and (b) to observe whether the identification of visemes is influenced by vowel context. There were no differences between the two groups either with respect to the overall percentage of items correct or to the visemes identified. Noticeable differences occurred in viseme identification between the /u/ context and the other 3 vowel contexts; visemes with /a/ differed slightly from those with / / and /i/; and there were no differences in viseme identification for / / and /i/ contexts. Findings were in general agreement with other studies with respect to the visemes identified, provided it is acknowledged that changes can occur depending on variables such as talkers, stimuli, recording and viewing conditions, training procedures, and statistical criteria. A composite grouping consists of /p,b,m/; /f,v/; /theta sigma/; /w,r/; /tS,d3S,3/; and /t,d,s,k,n,g,l/.
一系列由23个辅音与元音/a/、/i/、/u/和/ /构成的VCV无意义音节，以无声录像带的形式呈现给5名听力受损成年人和5名听力正常的成年人。双重目的是：（a）确定两组在视位识别方面的表现是否相同；（b）观察视位识别是否受元音语境的影响。两组在正确项目的总体百分比或所识别的视位方面均无差异。在/u/语境与其他3个元音语境之间的视位识别上出现了明显差异；/a/的视位与/ /和/i/的视位略有不同；/ /和/i/语境的视位识别没有差异。在所识别的视位方面，研究结果总体上与其他研究一致，但前提是要认识到，视位识别可能会因说话者、刺激、录制和观看条件、训练程序以及统计标准等变量而发生变化。一个复合分组包括/p,b,m/；/f,v/；/theta sigma/；/w,r/；/tS,d3S,3/；以及/t,d,s,k,n,g,l/。
